Recent evidence has implicated cell-cell interactions as a key element in directing cell fates in both vertebrate and invertebrate nervous systems. The Drosophila retina is no exception: several examples have now been defined in which a previously-differentiated cell provides an inductive cue to its uncommitted neighbor. However, this mechanism cannot account for specification of the first cell type. How is the chain of cell fate inductions begun? This proposal examines the mechanisms by which the first cell type, photoreceptor neuron R8, is specified in the developing Drosophila retina. I have recently conducted a detailed histological and genetic examination of early neuronal development. These studies have identified a group of 2- 3 cells which appear to be initially equipotent for the R8 fate, forming an "R8 equivalence group". This work suggests a simple model in which cells within this equivalence group interact to select a single R8 by stochastic means. An initial genetic screen has identified six mutations which appear to affect this process. Two approaches are proposed to further address specification of R8. The first approach involves identifying regulators of boss, a gene expressed early and exclusively in R8. The transcription factors Rough and Glass have already been demonstrated to play important roles in both boss regulation and R8 specification, suggesting this approach is a useful one. A series of promoter fusions and nested deletions are currently being tested to identify boss regulatory sites; a new technique has been developed which will reduce assay time from two months to five days. A second approach to identify regulators of R8 specification involves a genetic screen. Three genes which themselves were found to be required for R8 specification will be tested against a series of overlapping deficiencies to identify other interesting genes. To date, most of the molecules required for inductive interactions in the Drosophila retina have also been shown to play important roles in vertebrate- including human- development and oncogenesis. For example, R8 specification requires the homeobox protein Rough, a Drosophila homologue of a class of proteins which play a fundamental role in segmental identities within the immature mammalian nervous system; R8 specification also requires Drosophila homologues of the oncogene p2lras and its regulators. This project seeks to address how molecules and patterning cues combine to initiate cell identities within neuroepithelia, mechanisms which should also prove conserved between Drosophila and the significantly more complex mammalian nervous system.